notes from the use case breakout session
The task was to explore the use case analysis Rob suggested on Monday.
The group discussed several diverse use cases and attempted to decomplose them.
Determining essential elements and isolating common elements
The goal is to put together a reusable design pattern for approaching similar use cases
1. Virtual Organization with little structure in itself and no formal links between members in any single IdM system. Needs to interact w an org to manage grant money.
Ligo and NSF.
This one cascades into other use cases and needs more exploration
2. t nuclear lab
members of the VO coming from diverse locations need access to lab facilities at Duke U managed by differet groups at Duke. Many people in VO have no formal relationship w Duke
3. Univ libr has material in collection which it want s to make available to folks ... only when physically on univ property. Can be viewed by staff from anywhere
4. a member of VO wants access to a divers e set of wikis at a diverset set of institatuions . Wikis have links to each other. Person seeking material needs access that's transparent
Explored each in detail to understand the real world implications
Common factors: a VO with littler or no formal identitiy, needs to gain access to resources secured by a formal organ w strict accesscontrols
This process gave us fair confidence ....
We decided to start w a large set of use cases obtained from the wild??
More work than could be done in ½ hour.
Are there common elements in use cases? Or just coincidental similarities?
Need to validate that
Will comparisons results in so many different dimenseoins of comparisions aong different vectors that ...
Didn't get far enough with this, to see the circle in Rob's presentation of taxonomy use cases solution pattern, etc.
Made enough of a start to have good reason to be optimistaic.
Work should continue
Jon: U of Ill: Gang of 4 of design patterns. Big fran of that . concept of authorizaton speaks of this pattern concept. Is there any research or literature on this? Creating a library of auth patterns and relating them to solutions that work?
Could explore library space for work in this area.
Scott Fullerton: one of the things that struck me is we wrestled w one of the use cases trying determine if we were in the anaysiss space or the solution space.
The way we proceded varied as we tried to set that context
When we look for patterns, can collect patterns out in the wild
Another way is to look at what problems are they trying to solve and how would you ideally solve those problems. And what patterns would you find in those
There is a little difference in the approach
Need to find sepecific language, that's a primary aspect?